[unreadable] One of the experimental challenges in cancer molecular biology is assessing the validity and generality of biomarkers. This has become a critical bottleneck in the development of biomarkers from differential gene expression revealed by microarray studies. In this proposal, we develop the concept of using vertical arrays for exploration of differential gene expression in cancer. Vertical arrays explore the expression of a gene in many biological samples simultaneously, whereas standard microarrays explore the expression of many genes in response to one biological variable at a time. Vertical arrays are like dot blots in this regard, but vertical arrays are printed on glass slides, giving them better signal-to-noise behavior, and, rather than spotting the entire complexity of the RNA population in each spot, the RNA population is divided up among multiple spots. These low complexity representations have superb signal-to-noise performance. The work in this proposal will focus on establishing the feasibility of making a vertical array for studying gene regulation in many cancer samples simultaneously. Potential throughput is very high, such that multiple regions from each tumor can be studied simultaneously. This approach will be useful in confirming that a gene is indeed differentially regulated, in determining the distribution of expression of the gene in the transformed and surrounding normal tissue, and in determining whether the gene behaves in a similar manner in different cases of the same type of cancer and in different kinds of cancer. The goals require extensive and efficient microdissection, and we have built a novel instrument, the "tissue mill," to achieve these ends. Relevance: Biomarkers are useful for diagnosis, prognosis, and as potential therapeutic targets for cancer. There are hundreds of potential biomarkers, but further validation is needed before they can be exploited. [unreadable] [unreadable] [unreadable]